High frequency radio frequency (RF) communications are becoming more and more prevalent in the world today. Products touting wireless RF communication links are becoming increasingly popular among consumers. A multitude of new products including redesigned existing ones is being built with wireless RF links today. A RF link is based on a transmitter that emits a RF signal representing the information sent, and a communication receiver that receives the transmitted signal and extracts the information transmitted therefrom. To do this, modern RF receivers use an assortment of components such as amplifiers, filters, mixers, detectors, VCOs, PLLs, etc. A common goal in designing a receiver is to achieve high sensitivity and a low noise figure while maintaining interference rejection capabilities.
Typically, various types of potential interference must be considered, some being out of band signals, which may be suppressed by filtering means, and some being in-band signals which are not suppressed by the band filtering commonly placed at the input to the receiver, and therefore reach the receiver's active elements.
A block diagram illustrating a conventional prior art superheterodyne receiver front end is shown in FIG. 1. The receiver, generally referenced 10, comprises an antenna 12 coupled to the input of a band pass filter (BPF) 13. The output of the BPF 13 is then amplified by an RF amplifier 14. The RF amplifier is commonly constructed so as to be tunable over a range of frequencies. The output of the RF amp is then converted (heterodyned or mixed) by mixer 16. The mixer comprises a circuit that forms the product of two analog waveforms that contains the sum and difference frequencies of the signals at its inputs. The mixer thus functions to convert RF energy at one frequency to a second frequency. Mixers are commonly used in receiver front ends to convert input RF signal frequencies to lower intermediate frequencies. They are also, used in other circuit components such as upconverters, modulators, phase detectors, frequency synthesizers, etc.
The mixing action performed by a mixer is achieved by multiplying an input signal with a second signal, usually a local oscillator signal. The output comprises two signals at the sum and difference frequencies as shown below in Equation 1.
                              cos          ⁢                                          ⁢                      ω            1                    ⁢          t          ⁢                                          ⁢          cos          ⁢                                          ⁢                      ω            2                    ⁢          t                =                                            1              2                        ⁢            cos            ⁢                                                  ⁢                          (                                                ω                  1                                +                                  ω                  2                                            )                        ⁢                                                  ⁢            t                    +                                    1              2                        ⁢            cos            ⁢                                                  ⁢                          (                                                ω                  1                                -                                  ω                  1                                            )                        ⁢                                                  ⁢            t                                              (        1        )            One of the two output signals is the desired signal and is termed the intermediate frequency (IF). For example, the IF may be the difference frequency while the sum frequency is suppressed (i.e. filtered out) using a low pass filter. Additional frequencies, however, other than the sum and difference frequencies, are also generated by the mixer due to the use of nonlinear elements (e.g., diodes) to perform the multiplication. The nonlinearities of these components cause the generation of the additional frequencies.
A schematic diagram illustrating a typical prior art mixer circuit is shown in FIG. 2. The mixer circuit, generally referenced 30, comprises a mixer core made up of a differential amplifier including transistors Q1 and Q2. A local oscillator 34 drives the bases of transistors Q1 and Q2. Load resistors R1 and R2 are coupled between the supply voltage and the collectors of Q1 and Q2, respectively. The IF output signal is taken from the collectors of Q1 and Q2. The RF input signal is applied to the base of Q4 via coupling capacitor C1. The base of Q4 is biased by current source 32 in series with transistor Q3.
In operation, transistors Q3 and Q4 form a current mirror since the bases of both transistors are connected together. The circuit is configured such that Q1 and Q2 are biased in the nonlinear region at very low current near cutoff. For the mixer to operate, the current through Q1 and Q2 must be set accurately. Thus, Q1, Q2 and Q3, Q4 must be constructed as matched pairs of transistors.
Since the bases of Q3 and Q4 are tied together and the emitters of both are tied to ground, the current through one transistor must be duplicated in the other. This will be the case if both transistors are well matched since both will have the same VBE voltage drop. Consequently, the collector current will be the same through both transistors. To enable the mixing function, the constant current source 32 is configured to supply an amount of current such that both Q1 and Q2 are biased near cutoff.
In the absence of any RF input signal, the IF output signal is proportional to the local scillator signal 34. The gain of the output is set by the two load resistors R1, R2 When the
RF input signal is applied, the base voltage of Q3 and Q4 changes in accordance with the input signal. A rise in base voltage causes the current through Q3 and Q4 to increase. This causes the current flowing through Q1 and Q2 to increase. Transistors Q1 and Q2 now operate in a higher beta region. The change in the gain of the transistors causes a change in the IF output signal. Similarly, when the RF input decreases, the base voltage of Q3, Q4 also decreases. This causes a decrease in the current flowing through Q1 and Q2. Transistors Q1, Q2 now operate in a lower beta region causing a change in the output IF signal.
The nonlinear change in operating region, caused by the changes in the RF input signal, provides the mixing function whereby the desired sum and difference frequencies are generated. In addition, undesired intermodulation products are also produced. A narrow post mixer filter serves to remove most of the unwanted signals.
The generation of spurious output frequencies in a mixer is the result of using nonlinear switching elements to perform the multiplication function. Even when the input signal comprises a single frequency, the number of products generated may be large. The situation is compounded when the input signal comprises a plurality of frequencies as in various common scenarios of wireless reception.
A figure of merit that is indicative of the ability of a mixer to suppress intermodulation products is its third order intercept point, usually expressed in dBm. The intercept point is a measure of the linearity of a circuit or system permitting the calculation of distortion or intermodulation product levels from the amplitude of the input signal. An example graph illustrating the third order intercept for a mixer is shown in FIG. 3.
The desired signal component response (or 1st order linear response) is shown as line 40 (above the noise level 44) while the undesired 3rd order intermodulation components are shown in line 42. The two lines meet at a point 46 (usually beyond the 1 dB compression point). Assume two signals, having frequencies f1 and f2, are input to the receiver. The signal output of the mixer includes 3rd order interfering components (line 42) having frequencies 2 f1-1 f2 and 2 f2-1 f1. Note that these components are termed 3rd order due to the combination of a second harmonic and a fundamental. Since these frequency components are likely in the pass band, it is desirable to have as high a 3rd order intercept value as possible, which corresponds to high linearity and low levels of intermodulation products.
The theoretical input intercept point represents the input amplitude at which the desired signal components and the undesired signal components (i.e., third order distortion products) are equal in amplitude. Stated differently, the intercept point is arrived at by extrapolating measured data to yield an input RF level at which the IF level and intermodulation products would be equal.
The order of the intercept point determines the slope at which the amplitudes of the distortion products increase with an increase in the input level. For the case of the third intercept point (referred to as IP3), the intermodulation products increase in amplitude by 3 dB when the input signal is raised by 1 dB. The IP3 determines the amount of intermodulation distortion produced in the receiver itself when subjected to high-level interference. Note that a mixer having a high intercept point generates low intermodulation distortion products.
Further, the dynamic range of a receiver (i.e., the ability of the receiver to handle a wide range of levels of received signal and interference) is affected by the linearity of the RF receiver stages, such as the LNA, mixer, filters, detectors, etc. To improve the performance of the receiver, the dynamic range of the receiver must be widened by designing receiver stages having as high a 3rd order intercept point (IP3) as possible. In silicon integrated circuit (IC) design, however, a higher intercept point typically results in a higher Noise Figure (NF).
An amount of harmonics and intermodulation products are typically generated when two or more signals are input to the receiver due to the nonlinear elements in the receiver. It is desirable to design a receiver so as to minimize these products that degrade the receiver's performance.
The mixer that typically follows an LNA stage, having a gain in the order of 10 to 20 dB, is typically the component in the receiver most sensitive to nonlinearities. This is because it lies after the LNA wherein the signals are at a higher amplitude level having been amplified by the LNA. In addition, assuming a multi-channel FM receiver, any filters that lie before the mixer and/or the LNA must be fairly wide to accommodate the frequency range of channels the receiver must be able to handle. For example, many wireless systems operating in the 2.4 GHz ISM band must cover a frequency band of about 100 MHz. Such a wide input band permits not just the desired signal to enter the receiver but a large amount of undesired signals as well. Sources of undesired signals in the input frequency band of such a receiver include emissions from microwave ovens and various wireless transmissions in the 2.4 GHz ISM band. The fine tuning in such a receiver is not performed before the mixer but after it. The local oscillator signal input to the mixer is varied to correspond to the frequency of the desired received signal, while the channel selection filtering is performed by a fixed narrowband band pass filter after the mixer stage. Thus, the entire 100 MHz frequency band is constantly input to the mixer regardless of the desired received frequency.
The goal of the receiver design is to obtain the lowest noise figure possible while obtaining the highest 3rd order intercept possible. These goals, however, conflict with one another. To achieve a low noise figure, the gain of the LNA must be set high. In this case, large levels of noise and undesirable signals enter the mixer and discriminator due to the requirement of receiving all the channels in the band.
A common solution to this problem is to control the gain of the LNA thus adapting it to the level of the received desired signal and the relative level of interference present. A variable gain LNA is used typically having two gain states:
1. A high gain mode in which the LNA is active with a full gain of G dB, which is typically between 10 and 20 dB. The high gain of the LNA results in the lowest noise figure (NF) for the receiver.
2. A low gain setting in which the LNA is bypassed, thus reducing the total gain by G dB but substantially increasing the NF.
Note that in the more general case, the LNA may have an adjustable gain control which ranges between these two extreme gain values thus providing combinations of gain/linearity and NF that are in between the combinations of values presented above.
As in receivers based on discrete components, and in RF integrated circuit design, a higher 3rd order intercept point results in a high NF that could potentially degrade the performance of the receiver when weak signals are received. The first mixer state, which performs frequency downconversion in the front end of the receiver, typically follows a nominal LNA gain on the order of 15 dB. Therefore, in many cases the linearity of the receiver is limited by the 3rd order intercept point of the first mixer that must handle an amplified signal.
The selection whether to place the LNA in low or high gain mode should be determined in accordance with the received strength of the desired signal. In the case when the interfering signal is present and the desired signal is strong enough, switching the LNA to a low gain setting will reduce the effects of the interfering signal. If, however, the desired signal is weak and requires the LNA's amplification, the gain of the LNA must be kept high.
Setting the gain of the LNA to high gain mode places a considerable limitation on the linearity of the receiver. The limitation on linearity of the receiver is due to the low 3rd order intercept point of the first mixer stage. The reduction of LNA gain by 10 to 15 dB, intended to reduce the susceptibility of the receiver front end to nonlinear effects, causes an increase in the noise figure of the receiver and reduces its sensitivity. This is because the signals input to the mixer are not as strong as without the reduction in LNA gain. Consequently, the sensitivity of the receiver is reduced (the sensitivity is reduced as a consequence of an increase in the noise figure).
In some cases, it is not appropriate to switch the LNA to low gain mode. For example, in the case of near band high interference, the performance level of the receiver may fall below acceptable levels since the desired signal may be too low and the near band interference signal too high. In this case, the improvement of linearity is achieved at the expense of receiver sensitivity. Further, in some cases, switching the LNA to a low gain mode of operation does not result in a significant enough improvement of the linearity of the receiver.